A data-driven real-time evaluation method for safety risk of the anti-seepage system in high-core rockfill dams

Abstract

Risk evolution of the anti-seepage system in high-core rockfill dams is highly uncertain. Using traditional analysis methods to quantify the operational safety risks of dams is computationally intensive and lacks timeliness, making it difficult to achieve real-time risk evaluation. By investigating the nonlinear mapping relationship between the monitoring data (e.g., deformation, seepage pressure) and the time-varying risk, the surrogate models of time-varying risk with multiple failure modes are established. A coupled measuring point identification method for the different failure modes is proposed, and a monitoring data-driven real-time evaluation model is constructed, thus, forming a data-driven real-time evaluation method for the safety risk of the anti-seepage system in high-core rockfill dams that integrate the surrogate modeling of multimode risk, simulation of time-varying risk, identification of coupling measuring points, and fusion of monitoring information. The application shows that the proposed method has high accuracy and strong adaptability in predicting the system risk and different failure modes risk of the anti-seepage system, for example, the root mean square error is less than 0.07. The research findings provide a new approach for real-time evaluation of safety risks for large and complex hydraulic structures such as the anti-seepage system in high core rockfill dams.